‍

I went to law school because I wanted to help businesses navigate complicated legal situations. Yet, within weeks of arriving at Harvard Law, I learned that the legal profession was not what I had envisioned.

I learned that the practice of law for many lawyers was mostly document review, looking across thousands of pages of contracts and looking for abnormalities and for risk. In talking to attorneys in the field, it became obvious that no one enjoyed this type of work, especially not on a Friday afternoon.

I desperately hoped that there would be solutions out there to help with these challenges so I could focus on advising clients and working on the strategic components of lawyering. I couldn’t find a single solution that was widely adopted and being used effectively.

Along with my law school classmate and my client (who I met through a clinic and was a PhD candidate at MIT), we asked the question that very few lawyers asked at the time:

‍

Can we use technology to automate the boring parts of being a lawyer?

The answer was unclear in 2016. When we surveyed the landscape, we saw one of the most fragmented industries we’ve ever seen. Billing solutions, matter management platforms, contract generation companies.

We saw dozens of point solutions, which required clients to buy multiple pieces of software to piece together a holistic solution. From an AI perspective, most legal technology vendors were using rule-based approaches that had low accuracy rates.

‍

‍

The landscape on legal technology still is focused on software-based tools today.

We set out to build a system that was based on the most recent advances in artificial intelligence. The thinking behind our approach was that if a car can drive itself, we could pull out key terms out of a contract.

In the last 3 years, we’ve built a system that can accurately extract dozens of fields, both legal and financial, from even scanned third-party contracts.

We’re excited to share more about Evisort and our innovation on this blog.

‍

Written by Jerry Ting